Actionable graph insight
Consume the same package edges the CLI analyses. No heuristics—just real imports you can trace back to files.
Retracify
Dependency intelligence designed for modern engineering teams.
Retracify enumerates every package in your workspace, maps real import relationships, and produces artefacts you can drop into RFCs, CI, or architecture reviews without another dashboard.
- Setup
- Under five minutes locally
- Monorepos
- pnpm · npm · Yarn · aliases
- License
- MIT, open source
retracify report
workspace snapshot
42
packages analysed
0
critical blockers
132
module edges mapped
Key alerts
- @workspace/cart → @workspace/payments (cycle)
- @workspace/ui · 3 undeclared dependants
npx retracify .
# generates retracify.htmlBuilt for engineering conversations.
Retracify takes the graph your IDE hints at and turns it into precise, verifiable data. Swap hand-waving around “ownership” with concrete edges, usage counts, and external exposure that any engineer can follow.
Cycle & drift detection
Highlight strongly connected components, undeclared dependants, and dead installs before they surface as incidents.
Workspace native
Understands pnpm/npm/Yarn layouts, honours tsconfig aliases, and keeps nested packages isolated without noise.
Automation ready
Export JSON that mirrors the HTML payload so bots, scorecards, and CI gates operate on the same facts engineers see.
How teams put Retracify to work
01
Scan any repository
Point Retracify at the repo root. It discovers every package, respects ignore patterns, resolves aliases, and inspects actual import usage.
02
Review actionable insight
Cycles, undeclared usage, tooling footprint, and reference counts surface with the evidence required to justify refactors or guardrails.
03
Share or automate
Commit the HTML artefact alongside RFCs or wire the JSON into CI to flag cycles and undeclared dependants immediately.
Keep shipping fast without losing dependency discipline.
Retracify is open source, local-first, and tuned for engineers who want data they can trust. Run it whenever your dependency graph changes and keep the output close to the code.